Branch, M. (2014) ‘Malignant side effects of null hypothesis significance testing’, Theory & Psychology, 24 (2): 256–77.

This article argues that p-values provide no probabilistic information about the reliability of research findings and, furthermore, have a number of unwelcome side effects.

Munoz, A., Moreno, C. and Lujan, J., (2012) ‘Who is willing to pay for science? On the relationship between public perception of science and the attitude to public spending of science’, Public Understanding of Science, 21 (2): 242–53.

This study reports the results of a sample of nearly 7,000 respondents in Spain on attitudes to public spending on science and general perceptions of science. Its relevance to this chapter is that it compares a bivariate analysis between attitudes to public spending on science and a range of independent variables on perceptions of science using the statistic Cramer’s V with a multivariate analysis using discriminant analysis. The authors appear to suggest that the bivariate analysis was more helpful, but that none of the coefficients were vey high. The discriminant analysis (discriminant analysis is explained in Chapter 8 of Kent (2015)), however, did not enable the researchers to identify the characteristics that defined those in favour of public funding of science. This is precisely the kind of research that might have benefited from the configurational data analysis procedures that are explained in Chapter 7 of Kent (2015).